A blood test that predicts dementia 25 years before symptoms. A 4x risk multiplier for women on combined hormone therapy who happen to have elevated levels. Headlines that managed to scare both the women on HRT and the women considering it. Here is what the March 10, 2026 JAMA Network Open study actually says — and what it does not.
Key Takeaways
- Higher baseline p-tau217 was associated with 3x dementia risk and 2x mild cognitive impairment risk over 25 years
- Among women on combined estrogen-plus-progestin therapy, the dementia hazard ratio rose to 4.18 versus 3.07 in placebo
- Estrogen-only therapy did NOT show the same amplification
- Study used WHIMS data: women aged 65 to 79 starting therapy a decade or more after menopause
- p-tau217 is not yet a routine clinical test for HRT decisions
- The "critical window" hypothesis — start HRT within 10 years of menopause — is not contradicted by this finding
What the Study Actually Did
The cohort was drawn from the Women's Health Initiative Memory Study (WHIMS), which enrolled women aged 65 to 79 between 1996 and 1999. WHIMS randomized participants to either:
- Estrogen alone vs placebo (women without a uterus)
- Estrogen plus medroxyprogesterone acetate vs placebo (women with a uterus)
The researchers — led by groups at UC San Diego working with stored WHIMS samples — measured plasma p-tau217 from baseline blood draws and tracked the 2,766 participants for up to 25 years through 2021. The endpoints: who developed mild cognitive impairment (MCI), who developed dementia.
The Headline Numbers
| Outcome |
Hazard ratio per 1-SD increase in p-tau217 |
| MCI or dementia (combined) |
2.43 |
| MCI alone |
1.94 |
| Dementia alone |
3.17 |
A 1-standard-deviation increase in baseline p-tau217 was associated with about 3x the risk of dementia over the follow-up period. That is the main finding. It is a strong association.
The HRT Interaction
The interesting subgroup result was about interaction with hormone therapy.
- In women randomized to placebo, high p-tau217 was associated with 3.07x dementia risk
- In women randomized to estrogen plus progestin, high p-tau217 was associated with 4.18x dementia risk
- In women randomized to estrogen alone, no such amplification was observed — risk tracked the placebo group
The combined-therapy amplification is statistically meaningful and biologically interesting. It is also a subgroup analysis of a study that was not powered for this question, in a population that does not represent how most women take HRT today.
Why This Is Not the End of HRT
There are three reasons not to read this as a verdict against combined HRT.
Reason 1: WHIMS Enrolled the Wrong Window
WHIMS women started hormone therapy at ages 65 to 79. That is, on average, more than 15 years after their last menstrual period. The "critical window" or "timing hypothesis" of HRT and brain health says exactly this: starting hormone therapy more than 10 years after menopause looks bad in brain endpoints. Starting within 10 years looks neutral or favorable.
The KEEPS trial (Kronos Early Estrogen Prevention Study) and ELITE trial (Early versus Late Intervention Trial with Estradiol) both enrolled younger, recently-menopausal women and did not find adverse cognitive effects from HRT. The 2026 p-tau217 finding is consistent with the timing hypothesis. It is not a contradiction of it.
If you are within 10 years of menopause and considering HRT, this study is not your study. Our HRT brain fog and cognitive function article walks through the timing-hypothesis evidence in detail.
Reason 2: WHIMS Used the Wrong Hormones
WHIMS used:
- Conjugated equine estrogens (CEE) — derived from horse urine, a mix of estrogens including some not produced by the human body
- Medroxyprogesterone acetate (MPA) — a synthetic progestin with a different metabolic profile than the body's own progesterone
Most modern HRT protocols use:
- Estradiol — bioidentical, the human form of estrogen
- Micronized progesterone — bioidentical, the human form of progesterone
The 2026 study cannot tell us whether the dementia interaction would replicate with estradiol plus micronized progesterone. Some of the WHI cardiovascular harm signals turned out to be at least partly attributable to the specific drugs, not the hormones in general. The same caveat applies here.
Our bioidentical vs synthetic HRT for women article walks through the difference and why it matters.
Reason 3: Subgroup Analyses Are Hypothesis-Generating
The estrogen-plus-progestin amplification is a finding worth following up. It is not yet a finding worth changing protocols over. The study authors themselves call for replication in independent cohorts and in formulations that match modern practice. Subgroup interactions in observational follow-up of randomized trials have a track record of not always replicating.
What This Means If You Are On HRT Now
The practical guidance:
If You Are On Estrogen-Only HRT (After Hysterectomy)
This study did not see amplification of dementia risk in the estrogen-alone arm. Your regimen is not the one this paper is talking about. Stay the course unless your prescriber tells you otherwise.
If You Are On Combined HRT and Started Within 10 Years of Menopause
You are in the timing window where the broader evidence — KEEPS, ELITE, and observational analyses — supports HRT being neutral or possibly protective for cognition. The WHIMS-based finding is in a different population. Bring the study to your next visit. Discuss it. But do not unilaterally stop.
If you are on conjugated equine estrogen plus medroxyprogesterone acetate specifically — the WHIMS formulation — that is a fair conversation to have. Switching to bioidentical estradiol plus micronized progesterone is a reasonable option for many women based on the broader evidence base, even before this study.
If You Are On Combined HRT and Started After Age 65
Have the conversation. The WHIMS population was women starting HRT at exactly your stage. The dementia amplification with high p-tau217 was real in that group. Your prescriber should be weighing benefits (symptoms, bone protection, cardiovascular timing) against this signal.
If You Are Considering HRT Now
This paper does not change the standard advice: the best window for starting HRT — for symptoms, bone, heart, and probably brain — is within 10 years of menopause. Our best online HRT clinic review covers providers that do thorough cardiovascular and cognitive screening at intake and prescribe modern bioidentical formulations.
What About p-tau217 Testing Itself?
The p-tau217 blood test is increasingly available — labs including Quest, LabCorp, and several specialty Alzheimer's diagnostic services run the assay. Some are FDA-cleared as adjunctive diagnostic tools.
But: no major guideline body currently recommends p-tau217 screening before HRT initiation. The reasons:
- The study showing this interaction is single-cohort, not yet replicated
- The relevant population is older late-initiators, not the typical modern HRT candidate
- A positive p-tau217 in an asymptomatic woman has no clear actionable next step beyond "follow more closely"
- Cost is not trivial (often $200 to $600 for the cash-pay assay)
- Insurance coverage is inconsistent
If you are at high baseline genetic risk (APOE4 carrier, strong family history of early-onset Alzheimer's), the conversation with your prescriber may be different. p-tau217 testing in that subgroup is more defensible. But it is not yet standard.
Why APOE4 Status Comes Up Here
The 2026 study found that the p-tau217 association with cognitive decline was larger in APOE ε4 carriers — the genetic variant that confers increased Alzheimer's risk. This is consistent with what we know about APOE4 biology: carriers accumulate tau and amyloid more readily, and they appear to be the subgroup most affected by suboptimal hormone-therapy timing.
If you know you carry APOE ε4 (from 23andMe, a clinical test, or a research study), that is a fair conversation to have with your prescriber before starting combined HRT. It is not a contraindication. It is a reason to lean toward bioidentical formulations, the earliest possible start window, and possibly transdermal rather than oral estrogen delivery.
How This Fits With the Broader 2025-2026 HRT Landscape
The 2026 p-tau217 study lands during a major reset in how HRT is prescribed and discussed:
The p-tau217 finding is one more data point in this rapidly-moving landscape. It does not undo the case for HRT. It does sharpen the case for using modern formulations and starting in the right window.
What to Watch Next
Three follow-up questions need answers:
- Does the interaction replicate in modern formulations? Studies of estradiol plus micronized progesterone vs the WHIMS combination would be the obvious next step.
- Does the interaction hold for early initiators? Re-analysis of KEEPS and ELITE participants for p-tau217 status would test the timing hypothesis directly.
- Is p-tau217 testing actionable? A trial randomizing women with elevated p-tau217 to formulation A vs formulation B vs no HRT would tell us whether knowing your biomarker level should change the decision.
None of these are likely to land in 2026. Plan on a 2-3 year horizon for the replication evidence.
Bottom Line
The March 2026 JAMA Network Open study did three things:
- Confirmed that p-tau217 is a powerful predictor of dementia 25 years out — a finding that matters far beyond HRT
- Suggested that combined estrogen-plus-progestin therapy may amplify dementia risk in women who already have elevated p-tau217 at baseline
- Did NOT find the same signal for estrogen-only therapy
It does not say HRT causes dementia. It does not say women on HRT should stop. It does say the conversation about formulation choice and timing window matters more than ever — and that the population this study followed (WHIMS women starting late, on older formulations) is not the population most modern HRT candidates fit into.
If you are within 10 years of menopause and weighing HRT for symptoms, bone protection, or cardiovascular timing, the established benefit-risk math has not changed. Start the conversation with a clinic that uses modern bioidentical formulations and screens you properly. Our best online HRT clinic for women review covers which providers do that.
References
- Devick KL, Kawas CH, Manly JJ, et al. Plasma Phosphorylated Tau 217 and Incident Mild Cognitive Impairment and Dementia in Older Women. JAMA Network Open. Published online March 10, 2026. PMID: 41805953.
- Shumaker SA, Legault C, Rapp SR, et al. Estrogen plus progestin and the incidence of dementia and mild cognitive impairment in postmenopausal women: the Women's Health Initiative Memory Study. JAMA. 2003;289(20):2651-2662. PMID: 12771112.
- Henderson VW, St John JA, Hodis HN, et al. Cognitive effects of estradiol after menopause: A randomized trial of the timing hypothesis. Neurology. 2016;87(7):699-708. (ELITE trial cognitive outcomes.)
- Gleason CE, Dowling NM, Wharton W, et al. Effects of Hormone Therapy on Cognition and Mood in Recently Postmenopausal Women: Findings from the KEEPS-Cog Trial. PLOS Medicine. 2015;12(6):e1001833.
- Mosconi L, Berti V, Quinn C, et al. Sex differences in Alzheimer risk: Brain imaging of endocrine vs chronologic aging. Neurology. 2017;89(13):1382-1390.
- UC San Diego Health. "Blood Test Predicts Dementia in Women as Many as 25 Years Before Symptoms Begin." UCSD News. March 10, 2026.
- The Menopause Society. Position statement on hormone therapy. 2022 (current at time of 2026 study). https://menopause.org
Related Reading